It’s been a while since I did my last update. I thought I’d share some of my work thus far. I’ve spent the last couple of weeks gathering additional armour pieces so I can eventually have a massive spread of data for Determination information. I’m also working on obtaining more Bard, Ninja and Warrior FCOB armour pieces, so I can assess Job Coefficients more carefully and whether Jobs add an additional Coefficient to your Main Stat, Weapon Damage, or both (as a multiplier at the end).
Anyway, here’s what I’ve got so far.
This is a collection of Data Samples with varying values of Determination vs Strength. I’ve tested 421-320 Determination in most of the sets (harder to do so with 660 Strength as I’m lacking a few pieces of gear), but as you can see there’s a noticeable pattern. Similarly to the Graph Sunny Hirose made a while back (which is in one of my other Blog Posts), Damage Dealt becomes 0 once we’re around 1600-1700 Determination. The next thing is that with higher Strength, the scaling of Determination is much better than it would with lower Strength.
You can easily see this, as with “660 Strength”, or the blue line, we have a Determination Scaler of 0.1204, vs 0.1047 at 573 Strength.
So this partly answers one of my questions. Determination is a value which multiplies itself with Strength.
If I plot the all of the linear scalers into a graph, I get something like this:
With a Formula of STR*(0.0002*DET+0.0148). This only measures up Strength obviously. We still need to look at Weapon Damage… But I’m still collecting a lot more Data Before I begin to pursue that beast.
Respectively, we also have Auto-Attack Damage. As we all should know, Determination scales with Auto-Attacks completely differently, and this is evident by looking at this graph:
Now…. This graph is all over the place, but this is due to horrible rounding errors dealing with 57.22 Weapon Damage for my tests. The main thing we’ve gained from this however, is that the scaling for Determination vs Strength for Auto-Attacks is twice as much as Action Skills, a fact which Valky proved during Closed Beta 2.
Okay, so what now? I’ve got all of this Data, but what about a working Damage Formula? Well, I’ve actually gotten close, very close. I made a few different formula expressions, and ran my Data-Samples in a Linear scaling excel solver.. And so far, this my most accurate Expression:
(WD*A+B) * (STR*C+D) * (DET*E+F)
The Variance of Errors are on the Right hand Side. There’s a lot more data that isn’t shown, but the total Variance of Errors adds up to 4.020 and that’s with 64 Data-Samples. All of my other expressions have errors in the 20-100 range, so I think I’m rather close.
So, my two Final Formulas for the Dragoon so far are:
If I then create a graph comparing my data spread vs my damage expression, I get a graph like this:
Rather Accurate to my original Data Samples, no? All Damage meets at a mid section where I deal 0 damage equally, and stats scale exactly how they should be. I believe I may be one stop closer to my goal!
This formula obviously only works on the Dragoon and maybe the Bard, as each job affects Damage, or Stats, differently. If you want to see all of the work so far, below is my excel spreadsheet for all of my collected Data.